Fixed ambiguity between the row value syntax and INTERVAL syntax by making the

ROW keyword mandatory (although it's optional in SQL-99).


mysql-test/r/row_test.result:
  New results with the ROW syntax.
mysql-test/t/row_test.test:
  Updated tests to use the ROW keyword.
sql/sql_yacc.yy:
  Fixed ambiguity between the row value syntax and INTERVAL syntax by making the
  ROW keyword mandatory (although it's optional in SQL-99).
  The real problem might actually be INTERVAL, where the two rules
  "INTERVAL_SYM expr interval '+' expr" and
  "INTERVAL_SYM '(' expr ',' expr_list ')'" caused 40+ reduce/reduce conflicts
  with the rows value syntax.
  So possibly, the INTERVAL syntax should be fixed instead...

mysql-test/r/row_test.result

-SELECT (1,2,3)=(1,2,3);
-(1,2,3)=(1,2,3)
+SELECT ROW(1,2,3)=ROW(1,2,3);
+ROW(1,2,3)=ROW(1,2,3)
 1
-SELECT (2,2,3)=(1+1,2,3);
-(2,2,3)=(1+1,2,3)
+SELECT ROW(2,2,3)=ROW(1+1,2,3);
+ROW(2,2,3)=ROW(1+1,2,3)
 1
-SELECT (1,2,3)=(1+1,2,3);
-(1,2,3)=(1+1,2,3)
+SELECT ROW(1,2,3)=ROW(1+1,2,3);
+ROW(1,2,3)=ROW(1+1,2,3)
 0
-SELECT (1,2,3)<(1+1,2,3);
-(1,2,3)<(1+1,2,3)
+SELECT ROW(1,2,3)<ROW(1+1,2,3);
+ROW(1,2,3)<ROW(1+1,2,3)
 1
-SELECT (1,2,3)>(1+1,2,3);
-(1,2,3)>(1+1,2,3)
+SELECT ROW(1,2,3)>ROW(1+1,2,3);
+ROW(1,2,3)>ROW(1+1,2,3)
 0
-SELECT (1,2,3)<=(1+1,2,3);
-(1,2,3)<=(1+1,2,3)
+SELECT ROW(1,2,3)<=ROW(1+1,2,3);
+ROW(1,2,3)<=ROW(1+1,2,3)
 1
-SELECT (1,2,3)>=(1+1,2,3);
-(1,2,3)>=(1+1,2,3)
+SELECT ROW(1,2,3)>=ROW(1+1,2,3);
+ROW(1,2,3)>=ROW(1+1,2,3)
 0
-SELECT (1,2,3)<>(1+1,2,3);
-(1,2,3)<>(1+1,2,3)
+SELECT ROW(1,2,3)<>ROW(1+1,2,3);
+ROW(1,2,3)<>ROW(1+1,2,3)
 1
-SELECT (NULL,2,3)=(NULL,2,3);
-(NULL,2,3)=(NULL,2,3)
+SELECT ROW(NULL,2,3)=ROW(NULL,2,3);
+ROW(NULL,2,3)=ROW(NULL,2,3)
 NULL
-SELECT (NULL,2,3)<=>(NULL,2,3);
-(NULL,2,3)<=>(NULL,2,3)
+SELECT ROW(NULL,2,3)<=>ROW(NULL,2,3);
+ROW(NULL,2,3)<=>ROW(NULL,2,3)
 1
-SELECT (1,2,(3,4,5))=(1,2,(3,4,5));
-(1,2,(3,4,5))=(1,2,(3,4,5))
+SELECT ROW(1,2,ROW(3,4,5))=ROW(1,2,ROW(3,4,5));
+ROW(1,2,ROW(3,4,5))=ROW(1,2,ROW(3,4,5))
 1
-SELECT ('test',2,3.33)=('test',2,3.33);
-('test',2,3.33)=('test',2,3.33)
+SELECT ROW('test',2,3.33)=ROW('test',2,3.33);
+ROW('test',2,3.33)=ROW('test',2,3.33)
 1
-SELECT ('test',2,3.33)=('test',2,3.33,4);
+SELECT ROW('test',2,3.33)=ROW('test',2,3.33,4);
 Cardinality error (more/less than 3 columns)
 drop table if exists t1;
 create table t1 ( a int, b int, c int);
 insert into t1 values (1,2,3), (2,3,1), (3,2,1);
-select * from t1 where (1,2,3)=(a,b,c);
+select * from t1 where ROW(1,2,3)=ROW(a,b,c);
 a	b	c
 1	2	3
-select * from t1 where (0,2,3)=(a,b,c);
+select * from t1 where ROW(0,2,3)=ROW(a,b,c);
 a	b	c
-select * from t1 where (1,2,3)<(a,b,c);
+select * from t1 where ROW(1,2,3)<ROW(a,b,c);
 a	b	c
 2	3	1
 3	2	1
 drop table t1;
-select (1,1);
+select ROW(1,1);
 Cardinality error (more/less than 1 columns)
 drop table if exists t1;
 create table t1 (i int);
-select 1 from t1 where (1,1);
+select 1 from t1 where ROW(1,1);
 Cardinality error (more/less than 1 columns)
-select count(*) from t1 order by (1,1);
+select count(*) from t1 order by ROW(1,1);
 Cardinality error (more/less than 1 columns)
 drop table t1;

mysql-test/t/row_test.test

-SELECT (1,2,3)=(1,2,3);
-SELECT (2,2,3)=(1+1,2,3);
-SELECT (1,2,3)=(1+1,2,3);
-SELECT (1,2,3)<(1+1,2,3);
-SELECT (1,2,3)>(1+1,2,3);
-SELECT (1,2,3)<=(1+1,2,3);
-SELECT (1,2,3)>=(1+1,2,3);
-SELECT (1,2,3)<>(1+1,2,3);
-SELECT (NULL,2,3)=(NULL,2,3);
-SELECT (NULL,2,3)<=>(NULL,2,3);
-SELECT (1,2,(3,4,5))=(1,2,(3,4,5));
-SELECT ('test',2,3.33)=('test',2,3.33);
+SELECT ROW(1,2,3)=ROW(1,2,3);
+SELECT ROW(2,2,3)=ROW(1+1,2,3);
+SELECT ROW(1,2,3)=ROW(1+1,2,3);
+SELECT ROW(1,2,3)<ROW(1+1,2,3);
+SELECT ROW(1,2,3)>ROW(1+1,2,3);
+SELECT ROW(1,2,3)<=ROW(1+1,2,3);
+SELECT ROW(1,2,3)>=ROW(1+1,2,3);
+SELECT ROW(1,2,3)<>ROW(1+1,2,3);
+SELECT ROW(NULL,2,3)=ROW(NULL,2,3);
+SELECT ROW(NULL,2,3)<=>ROW(NULL,2,3);
+SELECT ROW(1,2,ROW(3,4,5))=ROW(1,2,ROW(3,4,5));
+SELECT ROW('test',2,3.33)=ROW('test',2,3.33);
 -- error 1239
-SELECT ('test',2,3.33)=('test',2,3.33,4);
+SELECT ROW('test',2,3.33)=ROW('test',2,3.33,4);
 drop table if exists t1;
 create table t1 ( a int, b int, c int);
 insert into t1 values (1,2,3), (2,3,1), (3,2,1);
-select * from t1 where (1,2,3)=(a,b,c);
-select * from t1 where (0,2,3)=(a,b,c);
-select * from t1 where (1,2,3)<(a,b,c);
+select * from t1 where ROW(1,2,3)=ROW(a,b,c);
+select * from t1 where ROW(0,2,3)=ROW(a,b,c);
+select * from t1 where ROW(1,2,3)<ROW(a,b,c);
 drop table t1;
 
 -- error 1239
-select (1,1);
+select ROW(1,1);
 drop table if exists t1;
 create table t1 (i int);
 -- error 1239
-select 1 from t1 where (1,1);
+select 1 from t1 where ROW(1,1);
 -- error 1239
-select count(*) from t1 order by (1,1);
+select count(*) from t1 order by ROW(1,1);
 #TODO remove comments after parser fixing
 #-- error 1239
 #select count(*) from t1 order by i having (1,1);

sql/sql_yacc.yy

@@ -1979,10 +1979,12 @@ simple_expr:
 	| NOT expr %prec NEG	{ $$= new Item_func_not($2); }
 	| '!' expr %prec NEG	{ $$= new Item_func_not($2); }
 	| '(' expr ')'		{ $$= $2; }
-	| '(' expr ',' expr_list ')'
+          /* Note: In SQL-99 "ROW" is optional, but not having it mandatory
+	     causes conflicts with the INTERVAL syntax. */
+	| ROW_SYM '(' expr ',' expr_list ')'
 	  {
-	    $4->push_front($2);
-	    $$= new Item_row(*$4);
+	    $5->push_front($3);
+	    $$= new Item_row(*$5);
 	  }
 	| EXISTS exists_subselect { $$= $2; }
 	| singleval_subselect   { $$= $1; }